CMOS image sensors may face new competition from an emerging imaging technology based on quantum dots.
Background: CMOS image sensors
Since the early 2000s, complementary metal-oxide-semiconductor (CMOS) image sensors have enabled smaller and cheaper digital cameras. Modern thin smartphones typically include multiple camera modules, eliminating the need for a separate camera for most users. However, conventional visible-light cameras still have limitations: highlights can saturate, low-light images can be noisy and lack vivid color, and infrared imaging typically requires different, more expensive detector materials with inferior image quality compared with visible-light sensors.
What are quantum dots?
Quantum dots are nanoscale semiconductor particles. Like other semiconductors, they release free electrons when they absorb photons. The key difference is that electrons in a quantum dot are confined by the particle boundaries because the dot diameter is only a few nanometers. This quantum confinement gives the particles distinctive properties.
One important property for imaging is tunable absorption: by selecting the material composition and particle size, manufacturers can tune which wavelengths quantum dots absorb, covering much of the visible and infrared spectrum. This tunability also applies to emission: recombining electrons emit light at a wavelength that can be precisely controlled. These emission properties have already been exploited in displays, where quantum dot enhancements are marketed under various names, commonly quantum dot light-emitting diodes (QLEDs).
Quantum dots are also compact and can be dispersed in printable inks, simplifying integration into manufacturing processes. Compared with silicon, quantum dots can absorb light more efficiently, enabling potentially thinner image sensor stacks. They also exhibit sensitivity across a wide dynamic range, from very low to very high illumination.
How quantum dot image sensors work
To understand quantum dot detectors, it helps to review how CMOS sensors convert light to images. In a typical camera, incoming light passes through optics and color filters before reaching silicon CMOS pixels. Filters determine which wavelengths each pixel records. When a silicon pixel absorbs a photon, an electron is freed and collected on an electrode or capacitor; readout circuitry converts the stored charge to a voltage that represents pixel brightness.
Conventional silicon detectors and readout circuits are fabricated together using well-established photolithography, etching, and deposition steps. In front-illuminated designs, metal contacts and interconnects above the detector reflect some incident light and reduce efficiency. Back-illuminated designs move the circuitry beneath the detector to avoid reflection losses, but this increases process complexity and cost. Silicon also becomes transparent beyond roughly 1 μm wavelength, so it cannot detect longer-wavelength near-infrared light.
Quantum dot detectors follow a similar readout scheme but replace the silicon absorber with a quantum dot layer. When quantum dots in a pixel absorb photons, electrons are released from localized bonds. If adjacent quantum dots are sufficiently close, electrons can hop between dots and reach the pixel electrode where the readout circuitry counts them. The readout circuits themselves can be fabricated on silicon wafers using standard processes.
Adding quantum dots to a wafer requires an additional processing step, but that step is comparatively simple: quantum dots can be suspended in an ink and printed or spin-coated onto the circuitry. This enables a back-illuminated pixel architecture without the added complexity and cost of conventional back-side processing, because nearly all incident light reaches the detector. Because quantum dots absorb light more strongly than silicon, a very thin layer above the readout circuit can capture most incident photons, which supports thinner sensors and improved dynamic range between low and high illumination.
Infrared imaging potential
Quantum dot detectors can be tuned into the infrared, which opens possibilities for more compact and lower-cost infrared cameras. Traditional infrared detectors use small-bandgap semiconductors such as lead sulfide, indium antimonide, mercury cadmium telluride, or indium gallium arsenide. These detector arrays are usually fabricated separately from silicon CMOS readout circuits and then hybridized: each pixel in the detector array is connected to the corresponding CMOS pixel through indium bumps and a bonding process that is time-consuming, limits array and pixel size, reduces throughput, and increases cost.
Quantum dots sensitive to the same infrared wavelengths can instead be synthesized via scalable chemical processes and then applied on top of completed silicon readout circuits. This removes the need for hybridization, allowing pixel sizes smaller than the typical ~15 μm indium-bump limit and enabling higher pixel counts in a smaller area. Smaller sensors reduce the size and cost of optics and can make infrared imaging more practical for a wider range of applications.
Challenges
Despite these advantages, quantum dot sensors face technical challenges before they can be widely commercialized. The main issues are stability, efficiency, and uniformity.
Quantum dots can oxidize in air, causing device degradation that manifests as reduced sensitivity, increased noise, slower response, or electrical shorts. In displays, encapsulation within polymer matrices and barrier layers is sufficient because quantum dots in display stacks do not need to transfer charge to external electrodes. In photodetectors, however, charge carriers must move through the quantum dot layer to electrodes, so simple encapsulation of individual dots within insulating polymers is not feasible.
One potential approach is to encapsulate the entire quantum dot layer or the entire device to allow charge transport while limiting atmospheric exposure. Alternatively, quantum dots can be chemically engineered to be less prone to oxidation while preserving charge transport properties and processability. Research on these approaches is ongoing.
Another challenge arises from organic ligands used to stabilize quantum dot surfaces. These ligands often act as insulators and impede charge transfer across the quantum dot film. Manufacturers commonly perform ligand-exchange processes to replace long insulating ligands with shorter conductive ones, improving film conductivity. However, this adds processing steps and can make quantum dots more susceptible to degradation over time because the exchange process may damage the quantum dot surface.
Efficiency is also a concern. Small quantum dots have a large surface-to-volume ratio and can host many defects; lattice defects can facilitate nonradiative recombination before photogenerated carriers reach the electrodes. Traditional single-crystal semiconductor detectors typically achieve quantum efficiencies above 50%, while many quantum dot photodetectors reported to date have efficiencies below 20%. Although quantum dots absorb light more effectively than silicon per unit thickness, their device-level efficiency has lagged due to recombination losses. Device design and material quality are steadily improving, and efficiencies have increased over time.
Finally, because quantum dots are synthesized chemically, some particle size variation is inherent. Optical and electronic properties depend strongly on dot size; variations in synthesis, purification, or storage can lead to batch-to-batch nonuniformity. Large, experienced manufacturers can better control size distribution, but smaller producers may struggle to maintain consistency.
Commercialization and outlook
Some companies are already working toward commercial quantum dot cameras. For example, SWIR Vision Systems introduced the Acuros camera, a short-wave infrared quantum dot-based device intended for applications where conventional infrared cameras are prohibitively expensive. The Acuros detector, based on lead sulfide quantum dots, currently achieves an average external quantum efficiency of about 15% in the targeted infrared band, well below the ~80% efficiency typical of indium gallium arsenide detectors. However, using 15 μm pixels, the Acuros camera can deliver higher resolution than many legacy infrared cameras, potentially making it attractive for maritime imaging, production inspection, and industrial process monitoring where cost matters.
In the consumer market, reports indicated that Apple acquired InVisage, a company developing quantum dot smartphone cameras. Apple has not publicly disclosed specific deployment plans, but infrared capabilities that offer lower cost and higher resolution for facial recognition and other sensing applications are likely of interest.
Other companies and research groups are actively addressing stability and efficiency limitations and exploring wavelength and sensitivity boundaries. Industry participants include BAE Systems, Brimrose, Episensors, and Voxtel, while academic research teams at institutions such as MIT, the University of Chicago, the University of Toronto, ETH Zurich, Sorbonne University, and City University of Hong Kong are contributing to the field.
It is plausible that within five years quantum dot image sensors could appear in smartphones, enabling improved low-light photography and video, enhanced facial recognition, and wider integration of infrared photodetection into consumer devices. Smaller, lower-cost sensors could accelerate adoption across many applications.
